It does not say "screw falling out completely would have been ok." The screw provided a single point of electrical contact for grounding. When it became loose (thus less surface area contact with the chassis and hence increased resistance in that part of the circuit) or fell out completely (zero surface area contact with chassis and hence infinite resistance in that part of the circuit) the current took a completely different path to the ground with catastrophic results.

I think this use of phrase "in fact" in the article is a bit Russian or otherwise Slavic. This usage shouldn't be illegal in English but is less common.

Electricity needs both positive and negative connections made between power source to equipment to work, but some old vehicles only ran the positive(sometimes only the negative) wires, and used the entire vehicle body the negative wire. Idea is that the electrons are gonna find its own way, the body's thick as it gets as a conductor, it's fine if you knew what you're doing, it saves lotta weight.

In this case, one of such connections to the body was secured with a screw, which could loosen from vibration, which `in fact` did. It instantly caused electricity to look for an alternative way, which was through propeller feathering switch and feathering relays, which caused the airplane to needlessly stop flying.

I think using "in fact" there is ok, I didn't have any trouble understanding it. "Actually" might have worked as well? And the part that follows, "with the grounding screw pulled out sufficiently far" probably also includes "the grounding screw missing completely".

Screw is forgotten or snapped: nothing is holding the wires, the resistance of ground connection is too high, switches either don't work or misbehave when you test them before flight. A lot of curse words about servicing the damn thing is heard.

Screw gets more and more loose (alternatively, dirt or rust accumulates): less and less current flows through the contact point, more and more voltage is applied to relays. Generally, it is not enough to switch them, until one day another bump changes the position of wires, and ads some more resistance…

If something fails, it should fail as a whole. It's analogous to aborting the program on unexpected error versus ignoring it and hoping that everything else generally works OK.

When the ground screw came loose the voltage coming out of the flap switch traveled back up the feathering ground, causing the system to think the feathering switch was activated.

When part of an electrical circuit starts to "float" like this - becoming no longer anchored to some global reference voltage (usually chassis ground is considered 0V) - the observed voltages in other parts of the system can take on really byzantine values.

Yes. It never said that the screw falling out would have been ok. The effect would have been the same where the flap ground connection would have found its way home via the feather system. Loose vs completely gone are just two points on a continuum of badness.